By tomski777 on

I'm currently working on a site which includes images and audio recordings. The site is part of an local history Project where we are collecting images taken in the last 100 years and recording interviews with local residents (greenwich London), many of whom have direct memories of the images being taken. These interview are to be transcribed, which has the handy effect of making the audio searchable :)

I've setup a taxonomy system so that images are being neatly oraganised and cross-referencing, using the following catagories:

  • Landmark
  • Company Name
  • Date Range
  • Subject Matter
  • Street Names

But I would now like to reference specific parts of the audio recordings so for example, all images tagged with a particular street, are also displayed with snippets of audio recordings refering to that same street etc.

So far, I have setup a local Darwin streaming server for testing and am able to pull out specific segmnents of the interviews using 'starttime' and 'endtime' comands in the embed tag (this site was really usefull for all this stuff: http://soundscreen.com ).

But now I'm wondering what the best way would be to link specific timecode to images. I could create another node type "streamed media snippet" or perhaps I could do something clever with the transcribed text (i'm not sure what!!) or perhaps I should use flash or something (though I'm anti closed formats).

Does anybody have any extra ideas/tips/references for how I could proceed?

Ta Tomski :)

⋅ Categories: Drupal 4.6.x

Comments

tannewt’s picture

tannewt commented

Tomski,
Check out annodexing. http://annodex.net It adds a cmml stream to ogg files.
~Scott (http://www.openroadtrip.net)

tomski777’s picture

tomski777 commented

Cheers for that tannewt - looks like a great project which I'll definitely look into more for another project on the horizon. & Sorry for the delay in replying, I just moved house & have been without a proper internet connection for f-ing ages!!

I ended up writing a small tagging module (a crude version of what Annodex is trying to achieve!) where the user creates a node which specifies start and end points within my streams (I was donated some space on a darwin server so used hinted .mp4 audio files), I then used a bit of javascript to play the audio at the required locations via an embeded quicktime ref file. As each reference is a node, this enabled me to tie all the audio into my taxonomy structure. Additionally I created (well its almost there;) a custom tag so users can embed specific sections of audio within any text on the site.

><>tomskii
><>www.theanthillsocial.co.uk